JPEG 2000 Encoding

You are here: JPEG 2000 Images > JPEG 2000 Encoding

JPEG 2000 Encoding

Output Format Specifier

Use the -outputformat/-of switch to specify the desired output JPEG 2000 format.

NOTE: GeoExpress supports encoding JPEG 2000 files in two different modes, "Part I" and "GMLJP2". GMLJP2 is the best option for most workflows. For applications that do not support GMLJP2, the Part I mode generates files that do not use any Part II extensions. GMLJP2 metadata is not available in this mode. Other encoding features, such as the use of NPJE profiles and other metadata forms, are still available. Use the Part I mode only when you are sure that the image user's workflow entails applications that cannot read GMLJP2 files.

The acceptable output format strings for JPEG 2000 encoding are:

gmljp2 – JPEG 2000

jp2 – JPEG 2000 (Part I)

nitfjp2 – NITF* encoded as JP2

The default output format is MG4.

*Support for encoding NITF images is available for GeoExpress. Contact your LizardTech representative to purchase the NITF Extension for GeoExpress.

The information included below is provided with the assumption that the user is familiar with the JPEG 2000 standard. Parameter choices accommodated by the JPEG 2000 specification are of such a broad and varied range that it is possible for users to select settings that prove counterproductive to their goals. In particular, some optimal encode settings can result in degraded decode performance. Considerable trial and error may be required to achieve desired results when diverging from defaults and recommended settings.

JPEG 2000 Encode Options

All of the basic encode options and some of the advanced options that are available for MrSID encoding are also available for the JPEG 2000 format. Certain advanced operations such as creating and updating composites and optimizing images are not available for JPEG 2000 encoding.

Options available for JPEG 2000 encoding include:

specifying a target compression ratio
encoding the image losslessly
selecting the number of zoom levels the output image has
specifying a target thumbnail size
encoding the image with the maximum number of supported zoom levels
creating a "test encode", which adds a watermark pattern and does not decrement the data cartridge

Color Space

JPEG 2000 images can be created with an essentially unlimited number of bands, so multispectral output can be created from multispectral input without having to go through an output color space. When outputting MrSID data, the color compositor is restricted to creating RGB color composites, but for JPEG 2000 encoding, the color compositor allows multibanded images to be created from files representing individual bands.

For more information see Creating Multiband JPEG 2000 Files.

NOTE: GeoExpress currently does not support the CMYK color space for JPEG 2000 images, either as input or output. For source images that are CMYK, the only supported output color space for JPEG 2000 encoding is grayscale. The preview input image color space controls remain the same.

Switches Used in Encoding JPEG 2000 Images

Basic JPEG 2000 Encoding Switches
Switch	Value	Notes
-profile	string	Specifies a predefined group of encode settings. Supported named profiles are: NPJE – “NGA-Preferred JPEG 2000 Encoding” settings. The NPJE profile sets the values to approximate the preferred parameters of the National Geospatial-Intelligence Agency (NGA) for encoding images for panning and zooming at high resolution. EPJE – “Exploitation Preferred JPEG 2000 Encoding” settings. The EPJE profile is optimal for low-resolution browsing among images. Large – Uses optimal settings for creating large JPEG 2000 images Default – LizardTech's general purpose JP2 encode settings. These are NOT the same as the GeoExpress default encode settings., which are recommended for large images.
-bitsofprecision/-bop	uint	Enables you to choose how many significant bits are to be used from the samples of the source imagery. For example, an image may consist of 16-bit data values, but only eleven of those bits may be meaningful. By default full precision is used.
-qualitylayers/-ql	uint	Indicates the number of quality layers with which to encode the image. Among other things, quality layers enable better “streaming” (progressive transmission and display) in some applications. Default: 30.
-tilesize/-ts	sint0 sint1	Specifies the use of tiles in encoding and their size (in pixels). If enabled, the tile size must be no larger than the dimensions of the source image. Smaller tiles reduce the amount of memory required to encode the image, however the use of tiles may introduce edge artifacts. Default: 4096 x 4096. Enter 0 x 0 to disable.
-flushperiod/-flush	uint	Enables codestream flushing, measured in rows of pixels. When codestream flushing is enabled, less memory may be required to encode the image. The flush period should be on the order of the strip height being used. Using the flush period should not significantly affect output image quality. Default: 4096.
-precincts/-prec	string	Specifies precinct sizes. Precincts are width-height pairs whose use may improve the performance of decode operations for certain image encodings such as those in which tiles are not used. The input format is `<w1`, `h1`, `w2`, `h2`,`...wn`, `hn>` where n is the number of zoom levels in the image. If you specify fewer than n precincts, the last one will be repeated for the remaining levels. Precinct sizes must be powers of two, except for the last one in the list, which has the lowest resolution. By default the use of precincts is enabled. Default for both width and height: 256.
-noprecincts/-noprec		Specifies disabling precincts.
-nogmljp2		Specifies that GMLJP2 metadata should not be included.
-noembedschemas		Specifies that GMLJP2 schemas should not be embedded in the output file.
-tileparts/-tp	string	Specifies the organization of tile parts in the codestream. Any combination of the letters "R", "L", and "C" is valid. N - disable tile parts R - group tiles by resolution L - group tiles by quality layer C - group tiles by component The default value is R.
-progressionorder/-porder	string	Progression order may have a significant impact on the time and memory usage required to encode and/or decode the image. Different progression orders should be used for different target workflows. Valid values are LRCP, RLCP, RPCL, PCRL, and CPRL. Default : RPCL.
-codeblocksize/-cbs	sint0 sint1	Specifies the width and height (in pixels) of the codeblocks used to encode the image. Default: 64. Note: for JP2 codeblock size restrictions, see JPEG 2000 Encoding.
-usepltmarkers/-plt		Enables the use of packet length markers. Use of packet length markers may improve decoding performance. By default packet length markers are used.
-use97wavelet		Specifies that the floating point 9-7 wavelet be used for encoding, instead of the default integer 5-3 wavelet. The 9-7 wavelet will not encode an image losslessly, however, it may result in a better looking image at higher compression ratios. This option is disabled by default and only available when the "`-lossless`" option is not specified.
-profilefile	string	Specifies path to a profile file to use for setting up encode properties. Additional properties set on the command line will override the profile settings.
-stripheight/-sh	sint	Affects memory usage. Increasing this value may improve runtime performance, but at a cost of requiring more physical memory. This value should be decreased when encoding very large images. Changing the strip height does not affect the quality of the output image. Acceptable values: Even integers from 2 – image’s height. Default: 12.
-tilelengthmarkers/-tlm	uint	Specifies use of tile length markers and their size. Using tile length markers may improve decoding performance with tiled JPEG 2000 files. Acceptable values: 1 – 255 Default: 255. Enter 0 to disable.
-nopltmarkers/-noplt		Disables packet length markers.

JP2 Codeblock Size Restrictions

The following restrictions apply to JP2 codeblock size:

width and height must be powers of two
width and height must each be greater than 4
width x height must be less than 4096
width and height must each be less than width and height of tile size
width and height must each be less than width and height of image if tile size is not specified

The default codeblock value is 64 x 64.

Mosaicking

Mosaic jobs are encoded the same way for JPEG 2000 output as for MrSID output files. See Creating a Mosaic.

JPEG 2000 Metadata

Evaluation Order of Metadata

If an external world file is present, its information trumps any metadata. Absent a world file, metadata are evaluated in the following order of priority:

GMLJP2
GEOJP2/GEOTIFF
MrSID metadata
World file metadata

NOTE: If a world file is present, then no coordinate reference system is associated with the output file (a NULL WKT is applied).

GMLJP2 Metadata

When JPEG 2000 is specified as the output format, GMLJP2 metadata is included by default, and GML metadata is added to the output file according to the GMLJP2 Application Schema version 3.1.1., which can be found at http://schemas.opengis.net/gml/ .

Embedding GML Schemas

If GMLJP2 metadata is included, the GML schemas are embedded in the file unless the -noembedschemas switch is used. XML data from the .xsd files that comprise the GMLJP2 Application Schema is embedded in the output file, eliminating potential dependency on a network for use of the GML data included in the file. The schemas package has a fixed size of 230 KB, which must be separately factored into output file size estimates (Note: this should only be of concern if the intended output imagery is under 10MB). If the -noembedschemas switch is used, software applications using the output file may need to retrieve the schemas online.

By default GeoExpress includes the schemas.

For more information about GML and the GML application schemas, see About GML and GMLJP2.

Default JP2 Settings

The table below indicates GeoExpress’ original or “factory” default JP2 settings. Unless you have changed your JPEG 2000 preferences, they are the settings that will serve as defaults for your JP2 encoding.

Default JP2 Settings
Option	Setting
Use of tiles	Enabled
Tile Size	4096 x 4096
Quality Layers	30
Strip Height	12
Flush Period	4096
Progression Order	RPCL
Code Block Size	Width: 64 Height: 64
Tile Length Markers	255
Tile Parts	By resolution
Packet Length Markers	Enabled
9-7 Wavelet (Lossy)	Disabled
Precincts	256 x 256
Metadata Forms Included in Output	MrSID, World file, GeoTIFF (GeoJP2)